- setup: switch to pyannote.core 2.0
- setup: switch to pyannote.core 1.3.2 / sortedcollectoins 1.0.1
- fix: fix corner case with less than 2 ORB descriptors
- chore: add FaceClustering at pyannote.video root level
- fix: add missing "--ffmpeg" option to command line tools
- improve: simplify installation (latest dlib, pip's opencv-python, & ffmpeg from imageio)
- fix: switch to pyannote.algorithms 0.8
- fix: fix shot theading
- doc: update "Getting started" guide
- fix: use latest pandas API (@crhan & @epratheeban)
- fix: use latest moviepy (@crhan)
- doc: update README
- setup: use latest pyannote.core
- feat: switch from Openface to dlib
- fix: fix Python 3 support
- doc: update 'Getting started' notebook
- fix: fix face clustering
- fix: fix support for latest numpy
- fix: fix Docker build
- setup: switch to pyannote.core 0.13
- setup: switch to pyannote.algorithms 0.7.3
- setup: switch to dlib 19.1.0
- fix: fix Docker build using latest pip version
- setup: switch to pyannote.algorithms 0.6.6
- fix: fix Docker build (@yaochx)
- fix: fix getting started guide (@yaochx)
- setup: switch to pyannote.core 0.6.6
- feat: add --torch parameter to pyannote-face.py script
- feat: pseudo-scenes segmentation based on shot threads
- doc: "getting started" notebook
- fix: include openface_server.lua in package
- fix: corner case in one-shot threading
- fix: remove hard-coded path to torch
- fix: corner case with iterable shorter than lookahead
- setup: update pyannote.algorithms dependency
- fix: include openface_server.lua in package
- fix: pyannote.core breaking changes
- feat: face clustering
- fix: OpenCV 2/3 support
- improve: shot threading
- improve: faster face tracking-by-detection
- fix: iter would incorrectly raise an IOError at the end of some Videos
- docker: pre-fetch MoviePy's own ffmpeg
- feat: face processing
- feat: shot threading
- feat: shot boundary detection (pyannote-shot)
- first public release